Materials Transactions, Vol. 43, No. 4 (2002) pp. 694 to 703
c
2002
The Japan Institute of Metals
Mössbauer Spectroscopic Study of Rust Formed on a Weathering Steel
and a Mild Steel Exposed for a Long Term in an Industrial Environment
Takayuki Kamimura1 , Saburo Nasu2 , Takashi Tazaki2, ∗ ,
Kaori Kuzushita2,∗ and Shotaro Morimoto2
1
Corporate Research and Development Laboratories, Sumitomo Metal Industries, Ltd., Amagasaki 660-0891, Japan
Division of Materials Physics, Department of Physical Science,
Graduate School of Engineering Science, Osaka University, Toyonaka 560-8531, Japan
2
The rusts formed on mild steel (15-year exposure) and weathering steel (32-year exposure) exposed to an industrial environment have
been characterized by means of X-ray diffraction technique and 57 Fe Mössbauer spectroscopy. By using an X-ray diffraction method, it is
suggested that the rusts formed on both steels consist of the crystalline α-FeOOH, γ -FeOOH and an X-ray amorphous phase, which gives no
peak to X-ray diffraction pattern. The amount of the X-ray amorphous phase exceeds 50% of the total amount of the rust. The 57 Fe Mössbauer
spectra observed at 10 K indicate that the rust contains only α-FeOOH, γ -FeOOH and Fe3−δ O4 (γ -Fe2 O3 ) for mild steel, and only α-FeOOH
and γ -FeOOH for weathering steel. The X-ray amorphous substance in the rust layer formed on mild steel possesses the structures of mainly
α-FeOOH showing superparamagnetism owing to its small particle size, and Fe3−δ O4 (γ -Fe2 O3 ). They are contained both in the inner rust
layer and in the outer rust layer. The X-ray amorphous phase in the rust layer formed on weathering steel is mainly α-FeOOH.
(Received November 16, 2001; Accepted February 19, 2002)
Keywords: weathering steel, mild steel, rust, goethite, lepidocrocite, iron oxyhydroxide, FeOOH, magnetite, maghemite, Mössbauer
spectroscopy
1. Introduction
An understanding of the mechanism of atmospheric corrosion of steel requires a study of the corrosion products (rust)
formed on the steel surfaces. It has been widely reported that
they consist of iron oxyhydroxides, such as α-FeOOH, βFeOOH, γ -FeOOH, and magnetite as crystalline substances,
and so-called “X-ray amorphous substance”.1–10) Their formation strongly depends upon the environment where steels
are exposed. β-FeOOH, for instance, is formed in the case of
existence of chloride.
Weathering steels, which contain a small amount of Cu, Cr,
P and Ni, form protective rust layers as corrosion proceeds
under atmospheric conditions. It is believed that the protective rust layer is composed of double layers of outer and inner
rusts, and the inner rust layer is dense enough to protect the
permeation of the corrosives. Under various environmental
conditions it is therefore of great importance to investigate
rust formed on steels because the formed rust layers, which
may act as coatings, greatly affect the corrosion resistance. To
investigate rust is also of great importance from a viewpoint
of electrochemical behavior of weathering steels, especially
under the initial wet/dry cycles. Because the rust is reduced
just after wetting of the dry surface as cathodic reaction by
the negative electrode potential, and during drying the oxygen reduction as cathodic reaction takes place on the reduced
rust surface in the pore system of the rust.11–22)
In spite of the importance of the rust on the corrosion
behavior of steels, limited numbers of studies have mentioned the rust constituents formed on the steel surfaces.
The crystalline rust, which is detected by X-ray diffraction
technique, has been well characterized, and the composition
change of the crystalline rust with the exposure duration have
∗ Graduate
Student, Osaka University.
been reported for the weathering steels.2, 3, 23, 24) In addition,
some different structures of the X-ray amorphous substance
have been proposed such as spinel type iron substance,4, 5)
Fe(OH)3 ,25) and δ-FeOOH10) by other authors. Recently, several pieces of literature on the investigation of the rust formed
on steels exposed to atmosphere have been reported by means
of Mössbauer spectroscopy26–31) because it has a great advantage for identification of poorly crystalline or amorphous corrosion products.32–37) Cook et al.26) and Yamashita et al.27)
pointed out that the X-ray amorphous substance formed on
the weathering steels exposed for 16 years in a rural environment having ISO site corrosion class C3 (S0, P0,T3)38)
is mainly Cr substituted α-FeOOH, of which crystal size is
less than 15 nm. Okada et al.39) have investigated the inner
rust, which was tightly adhered to the bulk steel, formed on
the weathering and mild steels exposed in a “semi-rural” type
of environment. They pointed out that α-FeOOH is the major component in the inner rust of both steels, and the difference in the structures of the inner rust between the mild
and the weathering steels is its particle size distribution; αFeOOH particle in the rust on weathering steel has a continuous distribution, while that on mild steel does not. The
above literature deals with the rust formed in the semi-rural
or rural environment. It is well known that the superior performance of weathering steel is generally obtained in the relatively severe conditions, where the pollutant such as SOx is
contained in the atmosphere; the great difference in corrosion
loss between weathering steel and mild steel is observed in
the industrial environment.40) We have investigated the rust
formed on the weathering steel in an industrial region for 15
years by means of Mössbauer spectroscopy and the following
results can be summarized.30) The X-ray amorphous phase
in the rust layer is mainly α-FeOOH which shows a paramagnetic doublet due to the superparamagnetism occurring
Mössbauer Spectroscopic Study of Rust Formed on a Weathering Steel and a Mild Steel
695
for small particles at room temperature. The outer layer also
contains a considerable amount of the superparamagnetic αFeOOH (spm. α-FeOOH). The particle size of α-FeOOH is
widely distributed and the inner layer contains relatively wellcrystallized α-FeOOH compared to the outer layer.
In the present study, to clarify the difference in the rust
constituents between those formed on the weathering steel
and those on the mild steel in an industrial environment,
where the weathering steel shows superior corrosion resistance to the mild steel, the constituents of corrosion products
formed on a mild steel exposed to an industrial environment
for 15 years are characterized by means of Mössbauer spectroscopy, and compared to the results previously obtained for
the rust formed on the weathering steel.30) Furthermore, the
rust formed on weathering steel exposed to an industrial environment for 32 years is also characterized in order to investigate the phase-change of the rust with the exposure duration.
were carried out according to the same procedure reported
previously.30) The cross-section of the surface rust layers was
investigated by the observation of the reflection behavior of
polarized light in an optical microscope using a Nikon Model
OPTIPHOT2-POL.
The 57 Fe Mössbauer measurements for the powdered rust
samples were performed in transmission geometry in the temperature range between 300 K and 5 K using a 57 Co γ -ray
source in Rh. The outer (10–20 µm in thickness) and the inner
(60–80 µm in thickness) rust were divided by the same procedure in the previous paper.30) The velocity scale is relative to
α-Fe at 300 K. Obtained spectra were analyzed by using a thin
foil approximation. Distribution analyses of the Mössbauer
parameters have been performed by using a NORMOS DIST
program developed by Brand.41)
2. Experimental
3.1 Constituent of rust
The surface of the mild steel exposed for 15 years and
the weathering steel exposed for 32 years were covered with
dark-brown rust, whose average thickness was approximately
60–90 µm. Figure 1 shows the cross-section of the rust. It
was found by microscopic observation, using reflected polarized light and crossed nicols, that the rust layer formed on the
weathering steel exposed for 32 years is mainly composed of
optically isotropic region (darkened), and only outer surface
of the rust is optically active (illuminated). On the other hand,
the rust formed on the mild steel is composed of the mottled
structure consisting of the optically active and isotropic corrosion products. For the mild steel the big cracks reaching
the steel substrate were clearly observed, whereas no crack
was observed for the weathering steel, which is the same tendency for the weathering steel exposed for 15 years.30) Those
cracks can be mostly formed during the sample preparation
of cutting and/or polishing procedure since they were not detected by the surface observation before cutting the samples.
It should be therefore noted that the rust formed on the mild
steel is fragile, and in other words, the rust formed on the
weathering steel can be adherent and tight.
From the X-ray diffraction patterns, the rust layers formed
on the mild steel and the weathering steel were mainly composed of α-FeOOH and γ -FeOOH as crystalline constituents,
and Fe3 O4 can be hardly detected in the rust of both steels
as shown in Fig. 2. The quantitative phase-analysis obtained
from the diffraction intensities indicated that the rust contained 20% of α-FeOOH, 15% of γ -FeOOH and 65% of the
X-ray amorphous substance for the mild steel, and 25% of αFeOOH, 10% of γ -FeOOH and 65% of the X-ray amorphous
2.1 Atmospheric rusting and preparation of specimens
Plates of mild steel and weathering steel (60×100×4 mm3 )
were exposed at Amagasaki, JAPAN at an angle of 30◦ facing south for 15 years and 32 years, respectively. This site,
located approximately 5 km inland from the coast but having below 0.01 mdd (mg/dm2 /day) of the deposition rate of
chloride, is considered industrial having ISO site corrosion
class C3 (S0, P1, T4). This site is identical to that in the
previous report.30) The chemical compositions of the steels
used for the exposure test and also used in the previous report30) are listed in Table 1. The specimens after the exposure
were cut perpendicular to the plate surface with a dimension
of 15×10 mm2 . A scrap of steel was first affixed to the surface
of the specimens with rust by hard adhesives (Torr Seal manufactured by varian) so that the rust on the steel is not scraped
off by subsequent polishing procedures, and the specimens
were then fixed by an epoxy resin. The cross-section of the
fixed specimen was mechanically polished by emery papers
(grades 400–1500) and then with a diamond paste. The rust
layers formed on the steel surface were also scraped off by
a razor until the steel surface appeared, and then they were
ground into powder. The powdered rust samples were desiccated for a week in advance of the subsequent analyses.
2.2 Analysis of the powdered rust and cross-section of
rust
The powdered samples were characterized by means of
X-ray diffraction technique (XRD). The XRD measurements
and the quantitative determination by peak intensities of rust
3. Results
Table 1 Chemical compositions of the steels tested (mass%).
C
Si
Mn
P
S
Cu
Ni
Cr
Mild steel
(15-years exposure)
0.12
0.32
1.47
0.023
0.004
0.01
0.02
0.04
Weathering steel30)
(15-years exposure)
0.11
0.24
0.75
0.014
0.005
0.33
0.12
0.49
Weathering steel
(32-years exposure)
0.10
0.50
0.50
0.018
0.014
0.55
0.53
0.91
696
T. Kamimura, S. Nasu, T. Tazaki, K. Kuzushita and S. Morimoto
Epoxy
Rust
(a)
100 m
Steel
Epoxy
Rust
100 m
(b)
Steel
ZnO (101)
Intensity (arbitrary unit)
-FeOOH
-FeOOH
ZnO (002)
ZnO (100)
Fig. 1 Cross-section of the rust formed on the weathering steel for 32 years (a) and the mild steel exposed for 15 years (b) by polarizing
microscope.
(a)
(b)
10
20
30
40
50
2
Fig. 2 Typical X-ray diffraction patterns of the rust formed on the steels.
(a) Weathering steel exposed for 32 years, (b) Mild steel exposed for 15
years.
substance for the weathering steel.
3.2 Mössbauer spectroscopy of the inner and the outer
rust
Figures 3 and 4 shows 57 Fe Mössbauer spectra obtained
from powdered samples scrapped from the inner and the outer
rust formed on the weathering steel exposed for 32 years, together with the results from the distribution analysis of the hyperfine magnetic field at 78 K at the right of the figures. The
line profile of the individual lines of sextets of the distributions was assumed to be Lorentzian and their width was fixed
to 0.4 mm s−1 for the fitting. The spectra obtained were similar to those for the weathering steel exposed for 15 years.30)
The room temperature spectrum of the inner rust shows a superposition of an intense quadrupole-split doublet and a magnetically split-sextet. This magnetically split-component is
divided into two sextets with strong and weak intensities.
Analysis was done by a least-square-fit with a thin foil approximation. The former sextet is clearly due to α-Fe having a magnetic hyperfine field Bint = 33 T at room temper-
ature, which was contaminated during the scraping of the
rust. The latter sextet is assigned to be α-FeOOH. The room
temperature spectra for the outer rust show only an intense
quadrupole-split. These results mean that the inner rust contains relatively well-crystallized α-FeOOH, and agree with
the results obtained for the rust formed on the weathering steel
exposed for 15 years.30) For the spectra of the inner and the
outer rusts at 78 K, the intensity of the central doublet decreased and the intensities of the sextets increased. These results suggest that the central paramagnetic component at 78 K
may be due to the paramagnetic γ -FeOOH whose Néel temperature is reported to be 73 K42) or 35 K43) and also due to
the spm. α-FeOOH26–28, 30) whose Néel temperature is much
higher than 78 K. This result is significant and leads to evidence of the existence of superparamagnetic small particles
in the rust layer formed on the weathering steel exposed for
32 years. As shown at the right in Figs. 3 and 4, the hyperfine
magnetic field of the sextet due to α-FeOOH is widely distributed at 78 K. These results may correspond to the wide distribution of the particle size of α-FeOOH. The central doublet
obtained at 78 K cannot be clearly identified because it may
contain smaller particles, which show paramagnetism even at
78 K. For the spectra of the inner and the outer rust at 5 K the
central doublet completely vanished, which implies that the
superparamagnetic fluctuation of the magnetic moments in a
nano-scaled particle of hydroxides may become static at 5 K.
The spectra observed at 5 K were analyzed into two sextets by
a least-square-fit; one is due to α-FeOOH having Bint = 50 T,
and the other is due to γ -FeOOH having Bint = 45 T. It should
be concluded that there is no phase-change of rust formed on
the weathering steels exposed for 15 years and 32 years.
Figures 5 and 6 show the 57 Fe Mössbauer spectra obtained
from powdered samples scrapped from the inner and the outer
rust formed on the mild steel, together with the results from
the distribution analysis of the hyperfine magnetic field at RT
and 78 K at the right of the figures, respectively. The spectra
obtained were different from that for the weathering steels.
The room temperature spectrum of the inner rust shows a
superposition of an intense quadrupole-split doublet and a
Mössbauer Spectroscopic Study of Rust Formed on a Weathering Steel and a Mild Steel
magnetically split-sextet. In addition, another broad sextet
was observed, which has a large full-width at half-maximum
(FWHM) and does not show the sharp spectrum. This sextet having a large hyperfine field of approximately 50 T was
never observed for the rust formed on the weathering steels.
The room temperature spectra for the outer rust also show an
intense quadrupole-split doublet and a broad sextet having a
large hyperfine field of approximately 50 T. For the spectra
of the inner and the outer rust layer at 78 K, the broad sextet observed at RT became sharpened, and the intensity of the
central doublet decreased and the intensities of the new sextet
lines increased. These results suggest that the central paramagnetic component at 78 K may be due to the paramagnetic
697
γ -FeOOH and the spm. α-FeOOH as mentioned above. This
leads to evidence of the existence of superparamagnetic small
particles in the rust layer for the rust formed on the mild steel.
Therefore the rust formed on the mild steel also contains a
considerable amount of the spm. α-FeOOH. As shown at
the right in Figs. 5 and 6, the hyperfine magnetic field of the
sextet due to α-FeOOH is widely distributed at 78 K. These
results may correspond to the wide distribution of the particle size of α-FeOOH, as mentioned above. The remarkable
differences in the distribution of the hyperfine magnetic field
were not observed between the mild steel and the weathering steel in this study. For the spectra of the inner and the
outer rust layer at 10 K the central doublet completely van-
Fig. 3 Typical transmission Mössbauer spectra of the inner rust formed on the weathering steel exposed for 32 years. The velocity is
relative to α-Fe. Right: hyperfine magnetic field distributions used to fit the spectra on the left.
698
T. Kamimura, S. Nasu, T. Tazaki, K. Kuzushita and S. Morimoto
ished. The spectra observed at 10 K were analyzed into three
sextets by a least-square-fit; one is due to γ -Fe2 O3 having
Bint = 52 T, another is due to α-FeOOH having Bint = 50 T,
and the other can be γ -FeOOH having Bint = 47 T, which
is slightly larger than Bint reported for the synthesized γ FeOOH42, 44) and the obtained γ -FeOOH in the rust formed
on the weathering steel as mentioned above. It is therefore
undeniable that a small amount of ferrihydrite may contribute
to this spectrum, whereas the Mössbauer spectrum of ferrihydrite is not yet clear.45–47)
4. Discussion
4.1 X-ray amorphous substance in rust formed on mild
steel
Magnetite (Fe3 O4 ) and magnetite (γ -Fe2 O3 ) were not observed in the rust formed on the mild steel used in this study
by means of XRD as shown in Fig. 2. The crystal structure of
these two iron oxides is cubic (inverse-spinel structure). The
lattice parameters are close to each other, and therefore they
cannot be distinguished easily only by XRD. The Mössbauer
spectra obtained for the rust formed on the mild steel clearly
Fig. 4 Typical transmission Mössbauer spectra of the outer rust formed on the weathering steel exposed for 32 years. The velocity is
relative to α-Fe. Right: hyperfine magnetic field distributions used to fit the spectra on the left.
Mössbauer Spectroscopic Study of Rust Formed on a Weathering Steel and a Mild Steel
shows the existence of 3rd substance except α-FeOOH and
γ -FeOOH. As has mentioned, α-FeOOH and γ -FeOOH were
detected by XRD as crystalline. From the hyperfine parameters of the 3rd substance at 10 K, it is likely that it is assigned
to γ -Fe2 O3 , because the spectra of the site for Fe(II + III)
of B-site in Fe3 O4 were not clearly observed at RT. For the
spectra observed at RT the broad sextet due to γ -Fe2 O3 is superposed on the intense doublet. This cannot be observed in
the rust formed on the weathering steel at all. It is proven
from the spectra at 10 K that there is no presence of intrinsic
amorphous phase in the rust. It should be pointed out that the
rust may contain a considerable amount of non-stoichiometric
substances. It is known that Fe3 O4 is partly oxidized and then
699
it can be expressed as Fe3−δ O4 (0 < δ < 1/3, δ = 1/3 corresponds to γ -Fe2 O3 .). Since the spectra from Fe(II + III) at
B-site in Fe3 O4 were not observed at RT as shown in Figs. 5
and 6, δ is thought to be mainly close to 1/3 (corresponding
to γ -Fe2 O3 ). However it is probable that the rust contains
Fe3−δ O4 having various values (0 < δ < 1/3). This may
result in the broad spectrum and widely distributed hyperfine
magnetic field at RT, shown in Figs. 5 and 6 on right hand
side. From the above-mentioned discussion, it may therefore
be concluded that one of the X-ray amorphous substance in
the rust formed on the mild steel is γ -Fe2 O3 (Fe3−δ O4 ).
The intensity of paramagnetic doublet is large at RT but
decreases at 78 K, and the contribution of the magnetically
Fig. 5 Typical transmission Mössbauer spectra of the inner rust formed on the mild steel exposed for 15 years. The velocity is relative
to α-Fe. Right: hyperfine magnetic field distributions used to fit the spectra on the left.
700
T. Kamimura, S. Nasu, T. Tazaki, K. Kuzushita and S. Morimoto
split-sextet, on the other hand, increases at 78 K. It should
be noted from these results that the rust formed on the mild
steel contains a considerable amount of spm. α-FeOOH. An
intense quadrupole-split doublet, except for γ -Fe2 O3 at RT,
is due to γ -FeOOH and the spm. α-FeOOH. It is therefore
concluded that the spm. α-FeOOH is also one of the X-ray
amorphous substances in the rust layer formed on the mild
steel. Furthermore, those γ -Fe2 O3 and the spm. α-FeOOH
are even contained in the outer rust layers.
Figure 7 shows the comparison of the mass fraction of α-FeOOH, γ -FeOOH and γ -Fe2 O3 in the rust
formed on the mild steel obtained by XRD and Mössbauer
spectroscopy at 10 K. Quantitative analysis of Mössbauer
spectra was performed by using the data of recoilfree fractions [ f (α-FeOOH)/ f (γ -FeOOH) = 0.98,
f (γ -Fe2 O3 )/ f (γ -FeOOH) = 0.92 at 77 K] reported by Oh
and Cook.48) The mass fraction of γ -FeOOH by different
quantitative determination is almost identical, and this also
supports that the X-ray amorphous substance is mainly the
spm. α-FeOOH and γ -Fe2 O3 in the rust formed on the mild
steel. The remarkable point obtained from Fig. 7 is that
approximately half of the X-ray amorphous substance is γ Fe2 O3 , and the residual is the spm. α-FeOOH. It is concluded
again that the X-ray amorphous substance in the rust formed
on the mild steel is composed of α-FeOOH and γ -Fe2 O3 although the rust may somewhat contain a small amount of γ -
Fig. 6 Typical transmission Mössbauer spectra of the outer rust formed on the mild steel exposed for 15 years. The velocity is relative
to α-Fe. Right: hyperfine magnetic field distributions used to fit the spectra on the left.
Mössbauer Spectroscopic Study of Rust Formed on a Weathering Steel and a Mild Steel
701
slightly increases with an increase of the exposure duration.
This tendency agrees with the results obtained by Yamashita
et al.,2) although their quantitative phase-determination was
done only by X-ray diffraction intensities. The increase in the
mass fraction of α-FeOOH in the rust with an increase of the
exposure duration was not observed for the weathering steels
exposed for 15 years30) and 32 years by X-ray diffraction
method, however the results by Mössbauer spectroscopy indicated the increase tendency of the mass fraction of α-FeOOH
with an increase of the exposure duration. The systematic investigation is needed to clarify the each mass fraction change
of rust with the exposure duration.49)
Fig. 7 Comparison of the mass fraction of rust obtained by X-ray diffraction and by Mössbauer spectroscopy for the rust formed on the mild steel.
Note that mass fraction of Mössbauer spectroscopy was calculated from
the data of the recoil-free fraction of γ -FeOOH, α-FeOOH and γ -Fe2 O3
by Oh and Cook,48) that is, f (α − FeOOH)/ f (γ − FeOOH) = 0.98 and
f (γ − Fe2 O3 )/ f (γ − FeOOH) = 0.92 at 77 K.
Fig. 8 Comparison of the mass fraction of the rust obtained by X-ray
diffraction and by Mössbauer spectroscopy for the rust formed on the
weathering steels. (a) Weathering steel exposed for 32 years [XRD],
(b) Weathering steel exposed for 32 years [Mössbauer spectroscopy], (c)
Weathering steel exposed for 15 years [Mössbauer spectroscopy]30) Note
that mass fraction of Mössbauer spectroscopy was calculated from the data
of the recoil-free fraction of γ -FeOOH and α-FeOOH by Oh and Cook,48)
that is, f (α − FeOOH)/ f (γ − FeOOH) = 0.98 at 77 K.
FeOOH, which can not be detected by means of XRD owing
to its small particles.
Figure 8 shows the comparison of the mass fraction of αFeOOH and γ -FeOOH in the rust formed on the weathering
steel obtained by XRD and Mössbauer spectroscopy at 5 K.
The mass fraction of γ -FeOOH by different quantitative determination is almost identical, and this also supports that the
X-ray amorphous substance is mainly the spm. α-FeOOH. It
is concluded that the X-ray amorphous substance is mainly
composed of α-FeOOH for the rust formed on the weathering
steel exposed for 32 years. The mass fraction of α-FeOOH
and γ -FeOOH for the weathering steel exposed for 32 years
is slightly different from that for 15 years,30) as shown in
Fig. 8(c). It is likely that the mass fraction of α-FeOOH
4.2 Difference between rust constituents formed on mild
and weathering steels
It was found that the X-ray amorphous substance in the
rust is α-FeOOH for the weathering steel and α-FeOOH and
Fe3−δ O4 (γ -Fe2 O3 ) for the mild steel. The exposure site is
identical, and therefore the difference of the rust constituents
is caused by the alloying elements such as Cu, Cr, Ni and
P. When one considers the process of the rust formation, the
solid state transformation of rust should also be taken into
account.50)
Misawa et al.7, 50) reported the detailed formation mechanism of iron oxides and oxyhydroxides. The fact that rust
formed on the mild steel contains Fe3−δ O4 may indicate one
possible rusting process, in which the rust can be formed via
green rust. Fe(OH)2 is first generated by the anodic dissolution during the atmospheric corrosion.51) It is gradually oxidized to green complex or green rust, which are the complex
compounds of Fe(II) and Fe(III). Fe(OH)2 has a hexagonal
close-packed structure of oxygen, and therefore it seems impossible that Fe3 O4 and γ -FeOOH is formed directly from
Fe(OH)2 because they have a cubic close-packed structure.
Therefore, Fe(OH)2 is first converted to the intermediate
phase (green rust) having both hexagonal and cubic layers of
closed packed oxygen, and then converted to Fe3 O4 and γ FeOOH. In the case of formation of Fe3 O4 , two oxygen ions
per three molecules of Fe(OH)2 must be removed with dehydration and deprotonation. If the oxidation proceeds slowly
and the oxidation time is long enough to remove oxygen from
crystal and rearrange oxygen and iron in crystal lattice, the
Fe3 O4 phase may be constructed. If the oxidation proceeds
fast and the structural rearrangement of green rust cannot be
completed, γ -FeOOH can be formed having the distorted cubic packing of oxygen. From the above-mentioned process of
Fe3 O4 and γ -FeOOH, the added alloying elements may affect the oxidation process of Fe(OH)2 . In other words, they
can affect the rearrangement of crystal structure. It was reported that Cu, which is one of the most effective elements on
the atmospheric corrosion of steel, can affect the oxidation of
the Fe(II)-complexes by catalytic effect,51) and therefore γ FeOOH can be formed without a formation of Fe3 O4 for the
rust formed on the weathering steel.
The other formation process of Fe3 O4 is the reduction
of the rust already formed on steels during the atmospheric
corrosion. The atmospheric corrosion is electrochemical in
nature.52) It is widely reported that rust is electrochemically
highly active,13–18) and therefore the rust on the corroding
steel surfaces is affected by the potential of steel. Due to the
702
T. Kamimura, S. Nasu, T. Tazaki, K. Kuzushita and S. Morimoto
variation in thickness of the electrolyte layer during exposure
to the atmosphere the rate of oxygen reduction should vary
periodically and this should result in a periodic variation of
the corrosion potential.
Stratmann et al.14, 18) have investigated the reduction behavior of rust in detail by means of in-situ Mössbauer spectroscopy. They demonstrated that the rust reduction behavior
is affected by pH, the concentration of Fe2+ and potential. αFeOOH is thermodynamically stable and not easily reduced,
although it is partly reduced at a very negative potential below
−0.5 V SHE. On the other hand, γ -FeOOH is easily reduced.
γ -FeOOH is reduced and {Fe.OH.OH} is formed probably on
the surface of γ -FeOOH as an intermediate.14) At the negative
potential such as −0.4 V SHE, which is nearly the corrosion
potential of Fe and, that is, the potential near the steel/rust
interface, γ -FeOOH is reduced to Fe3 O4 . They also demonstrated that the formed Fe3 O4 by the reduction of γ -FeOOH
is oxidized back not to γ -FeOOH but γ -Fe2 O3 . The added
elements are expected to affect the reduction behavior of γ FeOOH. In fact, Cu-incorporated γ -FeOOH is not easily reduced as compared to γ -FeOOH.53) Furthermore from the results of the potential measurements of the steels exposed for a
long period,54) the negative value of potential is observed for
the mild steel, while the more noble potential is observed for
the weathering steel because of the inhibition of the anodic reaction by the protective rust layer. This could result in the reduction of γ -FeOOH and the formation of Fe3 O4 , which can
be partly oxidized to γ -Fe2 O3 during a long period of exposure for the mild steel. From the results on the cross-section
of the rust as shown in Fig. 1, the rust formed on the mild
steel seems to be fragile. This may be related to the formation
of Fe3−δ O4 . If the formed γ -FeOOH is reduced to Fe3 O4 , the
volume change can be expected. If Fe3 O4 is formed and therefore the pore structure in the rust is changed, the subsequent
corrosion can be accelerated since Fe3 O4 having a good electric conductivity may act the cathodic site. From a viewpoint
of the reduction of rust, it is possible that even the rust formed
on weathering steels can also contain Fe3−δ O4 if the corrosion
potential stays at a very negative value; the thick water layer
forms repeatedly or for a long time on the steel surface, or the
atmosphere contains a large amount of air-borne salt particles
and protective rust cannot be formed.
It is still unclear whether Fe3 O4 and γ -FeOOH are formed
simultaneously at the rusting process or the initial corrosion
products is γ -FeOOH as has been stated by Misawa et al.,7)
and then the formed γ -FeOOH is partly reduced to Fe3 O4 after the subsequent corrosion process, whereas it is certain that
the added elements can be greatly related to both processes.
5. Conclusions
The rust layers formed on the mild steel (15-year exposure)
and the weathering steel (32-year exposure) exposed to an industrial environment have been characterized by means of Xray diffraction technique and Mössbauer spectroscopy. The
conclusions obtained are summarized as follows:
(1) The large cracks reaching the steel substrate are
clearly observed for the mild steel, whereas no crack is observed for the weathering steel. The rust formed on the weathering steel is presumably adherent and tight, and the outer rust
is also adherent and tight.
(2) The rust constituents formed on weathering steels are
not changed after a long exposure; they are only α-FeOOH
and γ -FeOOH.
(3) The X-ray amorphous substance in the rust formed on
the weathering steel exposed for 32 years is α-FeOOH having
the small particle size.
(4) The X-ray amorphous substance in the rust layer
formed on the mild steel, which gives no peak of X-ray
diffraction pattern, is mainly α-FeOOH showing superparamagnetism owing to its small particle size, and Fe3−δ O4 (γ Fe2 O3 ). They are contained both in the inner rust layer and in
the outer rust layer.
(5) Two possible formation processes of Fe3−δ O4 on the
mild steel are discussed. One is the oxidation of Fe(OH)2 via
green rust, and the other is the reduction of already formed
γ -FeOOH during the atmospheric corrosion of steel.
REFERENCES
1) T. Misawa, M. Yamashita, Y. Matsuda, H. Miyuki and H. Nagano: J.
Iron Steel Inst. Japan 79 (1993) 69–75.
2) M. Yamashita, H. Miyuki, T. Matsuda, H. Nagano and T. Misawa:
Corros. Sci. 36 (1994) 283–299.
3) M. Yamashita, H. Miyuki, H. Nagano and T. Misawa: Zairyo-to-Kankyo
43 (1994) 26–32.
4) H. Okada: J. Soc. Mater. Sci. Japan 17 (1968) 705–709.
5) H. Okada, Y. Hosoi, K. Yukawa and H. Naito: J. Iron Steel Inst. Japan
55 (1969) 355–365.
6) T. Misawa, K. Hashimoto and S. Shimodaira: Corros. Engineering
Japan 23 (1974) 17–27.
7) T. Misawa, K. Asami, K. Hashimoto and S. Shimodaira: Corros. Sci.
14 (1974) 279–289.
8) T. Misawa: Corros. Engineering Japan 32 (1983) 657–667.
9) H. Kihira, S. Ito and T. Murata: Corros. Sci. 31 (1990) 383–388.
10) J. T. Keiser and C. W. Brown: Corros. Sci. 23 (1983) 251–259.
11) M. Stratmann, K. Bohnenkamp and H.-J. Engell: Welkst. Korros. 34
(1983) 604–612.
12) M. Stratmann, K. Bohnenkamp and T. Ramchandran: Corros. Sci. 27
(1987) 905–926.
13) A. Kuch: Corros. Sci. 28 (1988) 221–231.
14) M. Stratmann, K. Bohnenkamp and H.-J. Engell: Corros. Sci. 23 (1983)
969–985.
15) U. R. Evans and C. A. J. Taylor: Corros. Sci. 12 (1972) 227–246.
16) I. Suzuki, Y. Hisamatsu and N. Masuko: J. Electrochem. Soc. 127 (1980)
2210–2215.
17) I. Suzuki, N. Masuko and Y. Hisamatsu: Corros. Engineering Japan 23
(1974) 141–149.
18) M. Stratmann and K. Hoffmann: Corros. Sci. 29 (1989) 1329–1352.
19) M. Stratmann and J. Müller: Corros. Sci. 36 (1994) 327–359.
20) M. Stratmann and H. Streckel: Corros. Sci. 30 (1990) 697–714.
21) T. Kamimura and M. Stratmann: Corros. Sci. 43 (2001) 429–446.
22) J. W. Schults, S. Mohr and M. M. Lohrengel: J. Electroanal. Chem. 154
(1983) 57–68.
23) T. Kamimura, M. Yamashita, H. Uchida and H. Miyuki: J. Japan. Inst.
Metal 65 (2001) 922–928.
24) M. Yamashita, A. Maeda, H. Uchida, T. Kamimura and H. Miyuki: J.
Japan. Inst. Metal 65 (2001) 967–971.
25) J. Dünwald and A. Otto: Corros. Sci. 29 (1989) 1167–1176.
26) D. C. Cook, R. Balasubramanian, S. J. Oh and M. Yamashita: Hyperfine
Interact. 122 (1999) 59–70.
27) M. Yamashita, T. Misawa, H. E. Townsend and D. C. Cook: J. Japan
Inst. Metals 64 (2000) 77–78.
28) M. Yamashita, T. Misawa, S. J. Oh, R. Balasubramanian and D. C.
Cook: Zairyo-to-Kankyo 49 (2000) 82–87.
29) J. F. Marco, M. Gracia, J. R. Gancedo, M. A. Martı́n-Luengo and G.
Joseph: Corros. Sci. 42 (2000) 753–771.
30) T. Kamimura and S. Nasu: Mater. Trans., JIM 41 (2000) 1208–1215.
31) S. Music, G. P. Santana, G. Šmit and V. K. Garg: Journal of Alloys and
Compounds 278 (1998) 291–301.
Mössbauer Spectroscopic Study of Rust Formed on a Weathering Steel and a Mild Steel
32) K. Nomura, M. Tasaka and Y. Ujihira: Corrosion 44 (1988) 131–135.
33) H. Leidheiser, Jr. and S. Music: Corros. Sci. 22 (1982) 1089–1096.
34) A. K. Sigh, T. Ericsson, L. Häggström and J. Gullman: Corros. Sci. 25
(1985) 931–945.
35) D. C. Cook: Hyperfine Interactions. 28 (1986) 891.
36) K. Nomura and Y. Ujihira: Hyperfine Interactions 29 (1986) 1467–
1470.
37) J. Davalos, M. Gracia, J. F. Marco and J. R. Gancedo: Hyperfine Interactions 69 (1991) 871–874.
38) International Standards Organization, “ Corrosion of Metals and Alloys
Corrosivity of Atmospheres Classification”, Standard 9223–92 (1992).
39) T. Okada, Y. Ishii, T. Mizoguchi, I. Tamura, Y. Kobayashi, Y. Takagi, S.
Suzuki, H. Kihira, M. Itou, A. Usami, K. Tanabe and K. Masuda: Jpn.
J. Appl. Phys. 39 (2000) 3382–3391.
40) For example, K. Horikawa: Kinzoku-Hyomen-Gijutsu 34 (1983) 529–
533.
41) R. A. Brand:
unpublished. Program is available through
http://www.wissel-instruments.de.
42) C. E. Johnson: J. Phys. C (Solid St. Phys.) 2 (1969) 1996–2002.
43) Ph. Bauer, J. M. Genin and D. Rezel: Hyperfine Interactions 28 (1986)
757–760.
44) S. Nasu, T. Kamimura and T. Tazaki: Hyperfine Interactions, to be pub-
703
lished.
45) E. Murad and J. H. Johnston: Iron oxides and oxyhydroxides, in: G. J.
Long (Ed.), Mössbauer Spectroscopy Applied to Inorganic Chemistry,
Vol. 2, (Plenum Publ. Co., New York-London, 1987) pp. 507–582.
46) J. F. Marco, M. Gracia, J. R. Gancedo, M. A. Martı́n-Luengo and G.
Joseph: Corros. Sci. 42 (2000) 753–771.
47) S. Nasu, T. Kamimura and T. Tazaki: unpublished data.
48) S. J. Oh and D. C. Cook: J. Appl. Phys. 85 (1999) 329–332.
49) T. Kamimura, T. Doi, T. Tazaki, K. Kuzushita, S. Morimoto and S. Nasu:
Proc. 2nd Int. Conf. on Environment Sensitive Cracking and Corrosion
Damage (2001) pp. 190–196.
50) T. Misawa, K. Hashimoto and S. Shimodaira: Corros. Sci. 14 (1974)
131–149.
51) T. Misawa, T. Kyuno, W. Suëaka and S. Shimodaira: Corros. Sci. 11
(1971) 35–48.
52) For example, M. Stratmann: Ber. Bunsen-ges. Phys. Chem. 94 (1990)
626–639.
53) M. Schneider and M. Stratmann: Ber. Bunsen-ges. Phys. Chem. 96
(1992) 1731–1736.
54) K. Kashima, S. Hara and H. Miyuki: Proc. 2nd Int. Conf. on Environment Sensitive Cracking and Corrosion Damage (2001) pp. 197–203.